[This article was first published on Fantasy Football Analytics » R | Fantasy Football Analytics, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)

---

Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

For a discussion of these metrics, see here and here.

The results are in the table below.  The rows represent the different sources of predictions (e.g., ESPN, CBS) and the columns represent the different measures of accuracy for the last two years and the average across years.  The “average” variable represents the mean of projections from ESPN, CBS, NFL.com, and FantasyPros.  The “latent” variable represents the common variance among projections from ESPN, CBS, NFL.com, and FantasyPros.  The source with the best measure for each metric is in blue.

Here is how the projections ranked over the last two years:

1. FantasyPros
2. Average
3. Latent
4. CBS
5. NFL.com
6. ESPN

Here is a scatterplot of the FantasyPros projections in relation to players’ actual fantasy points scored in 2013:

Interesting Observations

1. Projections that combined multiple sources of projections (FantasyPros, Average, Latent) were more accurate than single projections (CBS, NFL.com, ESPN).
2. The R-squared of the most accurate projection source was .67 in 2012 and .50 in 2013.  This suggests that players are more predictable in some years than others.  It also indicates that 1/3 to 1/2 of the variance in actual points is unexplained by projections, so there is much room for improvement in terms of prediction accuracy.
3. There was little consistency in performance across time among sites that used single projections (CBS, NFL.com, ESPN). In 2012, CBS was the most accurate single source of projection but they were the least accurate in 2013.  Moreover, the least accurate in 2012 was NFL.com, but they were the most accurate in 2013.  This suggests that no single source reliably outperforms the others.  While some sites may do better than others in any given year (because of fairly random variability–i.e., chance), it is unlikely that they will continue to outperform the other sites.  This may be similar to the finding that there is little consistency in the performance of mutual fund managers over time.  The following charts are from Leonard Mlodinow’s book, “The Drunkard’s Walk: How Randomness Rules Our Lives“.  The chart below depicts the performance of the top mutual funds from 1991 to 1995.

The next chart depicts the performance of the same funds in the same order over the subsequent 5 years (1996 to 2000).

In other words, the best funds from 1991-1995 weren’t necessarily the best funds from 1996-2000.  That’s likely why a cat beat investors in a stock market challenge.  This suggests that mutual fund managers differ in great part because of luck or chance rather than reliable skill.  Although our sample size is much smaller with fantasy football projections, there appears to be little consistency in fantasy football sites’ rank-order in accuracy over time.

Conclusion

In conclusion, the most accurate projections were from FantasyPros.com, which combines many sources of projections.  FantasyPros had the most accurate projections each of the last two years.  No single projection (CBS, NFL.com, ESPN) reliably outperformed the others, suggesting that differences between them are likely due in large part to chance.  I plan to calculate projections from more sites (Accuscore, Yahoo, FantasySharks) so we can compare even more projections.

---

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.

---

Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.